sNFl and sGFAP Predict MS Disability in Unique Ways

COPENHAGEN — Serum glial fibrillary acidic protein (sGFAP) is quickly maturing as a biomarker to predict disability in patients with multiple sclerosis (MS), but it will add information to, not compete with, serum neurofilament light chain (sNFl) levels, according to multiple independent studies.

The basic consensus is that “elevated sNFl levels predict inflammatory-associated worsening, while sGFAP values correlate with progression independent of inflammation,” said Enric Monreal, MD, Immunology Department, Ramón y Cajal University Hospital, Madrid, Spain.

This key message was repeated by several researchers presenting data at the 2024 ECTRIMS 2004 meeting, including one delivered as a latebreaker. There was also general agreement that sGFAP will eventually be a routine prognostic tool even if more data are needed to validate how it will be used in routine MS management.

A New Biomarker for MS Disability Progression

Although apparently reliable for predicting MS disability, “sGFAP is about 5 years behind where we are with sNFl,” said Evan Madill, MD, a clinical research fellow at the Brigham Multiple Sclerosis Research Center, Harvard Medical School, Boston. He does think, however, that it is coming to clinical practice.

In the study he presented, 744 patients from the Brigham MS Research Center database were evaluated retrospectively for sGFAP levels and subsequent disability progression. Among this cohort, for which sGFAP levels were collected at baseline and over time, 46.5% had 6-month confirmed disability progression (CDP) over follow-up.

On univariate analysis, sGFAP levels correlated with and predicted CDP, need for a new ambulatory aid, and conversion to secondary progressive MS (SPMS). For patients less than 60 years of age, all of these correlations were highly significant (P ≤.002). On multivariate analysis, the significance was preserved for CDP (P =.032) and for need of a new ambulatory aid (P =.007), but it was lost for SPMS conversion.

Notably, his data suggest that a one-time baseline measurement of sGFAP was more useful than change in sGFAP as a predictor.

It is unclear why sGFAP is less predictive in older individuals, but Dr Madill speculated that non-MS phenomena might play a role at older ages. Treatment did not influence sGFAP levels in this study, but Dr Madill said most of the data were collected before anti-CD20 monoclonal antibodies were widely available.

The observational study data presented by Dr Monreal involved 725 patients drawn from 13 European hospitals. sGFAP and sNFl levels were evaluated from blood drawn within 12 months of MS onset. Over time these biomarkers had overlapping but different predictive strengths.

Consistent with previously published studies, which link elevations in sNFl to neuronal damage and elevations in sGFAP to astrogliosis, sGFAP was found to be more useful for predicting progression independent of relapse activity (PIRA), particularly in patients with low sNFl levels.

Increases in sNFl were associated with an increased risk of both PIRA and relapse-associated worsening (RAW), but sNFl was more closely associated with RAW in untreated patients. The risk of PIRA and RAW were similar across GFAP and sNFl levels in those patients treated with high-efficacy disease-modifying therapies (DMT).

Overall, when stratifying the cohort into three groups, those with both low sNFl and low GFAP, those with high sNFl with low GFAP, and those with high GFAP and low sNFl, the relative risks of disability associated with PIRA and RAW diverged, suggesting these biomarkers correlate with different processes of progression.

Comparing sGFAP and sNFl

This same principle was explored further in the latebreaking presentation by Ahmed Abdelhak, MD, a clinical instructor, Weill Institute for Neurosciences, University of California, San Francisco. The objective of his study was to compare sGFAP and sNFl for predicting PIRA in patients on treatment.

The study included 212 patients from the Swiss Multiple Sclerosis Cohort who were started on fingolimod or on B-cell depleting therapies like rituximab. After correcting for sex, age at onset, baseline Expanded Disability Status Scale (EDSS) scores, and other variables, Dr Abdelhak also reported that the predictive values for PIRA were different for sGFAP relative to sNFl at least on the group level.

However, in this study, unlike the analysis of the Brigham MS Research Center data, changes in sGFAP over time when on treatment did have prognostic value, and there was a relationship between sGFAP levels and treatment. Although reductions in GFAP predicted less disability progression whether patients were treated with fingolimod B-cell depleting therapies, that patterns were different. Dr Abdelhak, like the other investigators speaking at ECTRIMS, also said the data so far favor sGFAP over sNFl for predicting PIRA.

Each z-score unit change in sGFAP corresponded to a 47% lower risk of PIRA in follow-up over 6.8 years, Dr Abdelhak reported, adding that the predictive value of sGFAP was “numerically stronger than the corresponding relation for sNFl.”

So far, clinical utility of sGFAP remains speculative. Most of the correlations he presented were on a group rather than the individual level. Moreover, Dr Abdelhak cautioned that these correlations, based on observational data, do not necessarily reflect causation.

Nonetheless, remarking on the parallels of his data on sGFAP and sNFl with other studies presented at the ECTRIMS meeting, Dr Abdelhak foresees a time when GFAP will be a prognostic tool, offering relative simplicity and lower cost than the current standard of imaging. He also sees a role in clinical research.

“Monitoring of sGFAP dynamics following DMT initiation could be used to prognosticate long-term PIRA risk and provide insights valuable for design and interpretation of trial outcomes,” he said.

Dr Monreal reported financial relationships with Almirall, Biogen, Bristol Myers Squibb, Janssen, Merck, Novartis, Roche, and Sanofi. Dr Madill and Dr Abdelhak reported no potential conflicts of interest. 

This article originally appeared on MDedge.com, part of the Medscape Professional Network.